ISM Book Chapter 2 - performance and IOPS

Hey all,

I was going over chapter 2 in the ISM book once more, and I continue to have some trouble with chapter 2 of the book. In every other chapter you find explanations for things, but with the formulas (Little's law, utilization law, service time, etc) you get a lot of condensed info at once and I found (and still find) this one of the hardest parts to congest in the book.

So, to see if I actually understood the chapter, I'm just picking a random drive and crunched some numbers here to see if I actually understood everything, this might help others, and if I made a mistake somewhere other might be able to help me?

So, as a reference drive, I took the Seagate Cheatah 15K.6 300GB drive with a FC interface (Model ST3450856FC). Data spec sheet can be found here.

Specs:

Model Number	ST3450856FC
Interface	Fibre Channel 4 Gb/s
Cache	16MB
Capacity	450 GB
Areal density (avg)	165 Gbits/inch2
Guaranteed Sectors	879,097,968
Spindle Speed	15,000 rpm
Sustained data transfer rate	171,000 null
Average latency	2,0 msec
Random read seek time	3,4 msec
Random write seek time	3,9 msec

Looks like this should contain everything we need to do the calculation. Let's also assume we are connected to a SCSI3 U320 controller and running 4 disks in a raid 1+0 configuration. We have a server set up with a DB that has block sizes ranging from 4KB up to 64KB

The problem I face each time is that some of the formulas explained in the book seem to depend on other formulas, but the explanation where I actually get this information from is a bit unclear to me. It talks about the service time but never states how to calculate the service time, or where to get the service time from. Perhaps I overlooked it or perhaps it might be a good idea to provide some more explanation to these equations in a new print of the book?

As far as I came with my number crunching I got to around 250 IOPS for this drive. Now for one I am not certain that this is correct, and second of all somehow my calculations seem to vary a bit.

Can anyone tell me if this is correct, and perhaps explain how to calculate this number a bit more clearly than in the book?